Document Memory Builder

Quick start

1. Require the document set and intended output location before doing substantive work.
2. Ask for the downstream purpose only if it materially changes what should be remembered; otherwise default to general reusable project memory.
3. Inventory the sources, identify canonical documents, and note recency/version signals.
4. Run python3 scripts/init_memory_pack.py <memory-name> --path <output-dir> to create the memory pack scaffold.
5. Read the corpus in priority order, extract durable information, and keep every memory item traceable to one or more sources.
6. Write the pack, explicitly separating stable memory from volatile items and unresolved conflicts.

Inputs and defaults

Required

• documents: explicit file paths, URLs, or a clearly bounded document set.
• output_dir: where the memory pack should be written.

Optional